Closed smiths closed 2 years ago
@smiths This is addressed in b987caec9693432a11715911f528b4a97cae0fb0.
I have updated the measurement spreadsheet where appropriate, but some of this information was never part of those empirical measures. I have added another spreadsheet (Artifacts.xlsx) that I used to collect data regarding this issue.
In summary:
3 projects (ESPREsSo, Ludwig, Palabos) had a decent contributor document at the time of data gathering (2020). One project (ESPResSo++) had a good contributor wiki page at that time. Eight other projects have either incomplete contributor documents (for example a contributor document that briefly comments on how to use git, but does not comment on the actual project) or have brief comments on contributing within other artifacts.
10 projects comment on citation. OpenLB has this on a separate webpage. Most of the projects that comment on citation have this information as part of their main README file.
Only one project (pyLBM) has an uninstall feature. In this case this is a feature/option of the software. This is not mentioned in the repo or in accompanying documentation. This is not an artifact. I did not find any uninstall artifact in the repos.
Three projects (DL_MESO, MechSys, Sailfish) have acknowledgements as part of other artifacts/webpages.
Three projects (HemeLB, LUMA, Musubi) have code style artifacts.
Espresso++ has a roadmap GitHub wiki. It is the only project with a roadmap artifact. Roadmaps were mentioned in comments elsewhere.
20 projects have some sort of getting started artifact or subsection of an artifact, sometimes related to tutorials.
It looks like most artifacts will at least be marked as rare in Table 8. The only concern is uninstall artifacts.
Great! This is very useful data. I'm not sure why you had to create a new spreadsheet though. Wasn't there a cell in the original spreadsheet where artifacts were listed? To record this information, we can just add these artifacts to the list that cell.
@smiths, I have added this info into the relevant cell of the empirical measures spreadsheet in f78677a38dd45d892369ff0ffa268a9ef585207a.
That empirical measures spreadsheet also lists artifacts other than those addressed in this issue in each project's relevant cell. The new spreadsheet will make things easier to read when addressing this issue, and it can can be deleted when it is no longer needed.
I also want to bring your attention to comments that I noticed yesterday in the DL_MESO manual regarding citation of the project:
4.2 The Licensee will ensure that, if any of its employees or
students publishes any article or other material resulting from,
or relating to, a project or work undertaken with the assistance
of any part of the DL_MESO Software, that publication will
contain the following acknowledgement:
"DL_MESO is a mesoscale simulation package written by R. Qin,
W. Smith and M. A. Seaton and has been obtained from UKRI
STFC’s Daresbury Laboratory via the website
http://www.ccp5.ac.uk/DL_MESO"
and cite the following reference:
"M.A. Seaton, R.L. Anderson, S. Metz & W. Smith, ‘Mol. Sim.’,
39 (10), 796-821 (2013)".
According to the manual we must cite the project in such a way.
I updated the SOP for LBM paper with the new artifact count data in 39bc3ab.
With respect to the DL_MESO, I think we are fine with what we currently have. We have cited the paper that is mentioned. The other part of the acknowledgement is for people that use DL_MESO to do simulations. We didn't really create our work with the assistance of DL_MESO.
@peter-michalski, I would like to add a figure (inspired by our ICCS paper) to the SOP for LBM paper. The figure maps the artifacts to software guidelines and to LBM software. A draft of the figure is found in Table 8 for commit #94daa95. I haven't yet filled in the last column showing which artifacts appear in LBM software. (The entries shown are actually for MI software.) As it is right now, we could leave several artifacts blank, implying that they don't appear in LBM software. My intuition is that they do appear, but since we weren't looking for them, they weren't noted. This was my experience with the MI software.
Can you please go through the LBM software packages again and count how many show each of the following artifacts:
The count might be zero for some of these, but I feel like we should give it one more look, since our results can be criticized if someone finds one project with a CONTRIBUTING file.
If you find any of the above artifacts, please update the measurement spreadsheet with the results. I'll make sure to update the Mendeley versions of the spreadsheets to have the most current data.